Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method for high-dynamic range (HDR) imaging, comprising: identifying one or more faces in a preview image; identifying a region of interest (ROI) for each identified face of the one or more faces in the preview image, wherein a number of identified ROIs is more than one; determining a number of images to be captured for an HDR image to be the number of identified ROIs; and for each ROI corresponding to an identified face of the one or more faces: determining a skin tone of the identified face; determining, based on the determined skin tone and a comparison between the identified face and the ROI corresponding to the identified face, a target brightness of the respective image associated with the ROI; and determining, based on the target brightness, an exposure value of the respective image associated with the ROI.
This invention relates to digital imaging, specifically to improving the quality of high-dynamic range (HDR) images captured in challenging lighting conditions. The problem addressed is the difficulty in accurately capturing detail across a wide range of brightness levels, particularly when faces are present in the scene. The method involves analyzing a preview image to detect one or more faces. For each detected face, a specific region of interest (ROI) is defined within the preview image. The total number of these ROIs dictates the number of individual images that will be captured to construct the final HDR image. For each identified face and its corresponding ROI, the method proceeds to determine the skin tone of the face. Based on this skin tone and a comparison between the face and its ROI, a specific target brightness is calculated for the image that will be associated with that ROI. Finally, an appropriate exposure value is determined for each individual image based on its calculated target brightness, ensuring optimal detail capture for each facial region within the HDR composition.
2. The method of claim 1 , wherein determining the exposure value of a respective image associated with a ROI comprises exclusively using the ROI in determining the exposure value.
This invention relates to image processing, specifically to methods for determining exposure values in digital imaging systems. The problem addressed is the need for accurate exposure value calculation, particularly when focusing on regions of interest (ROIs) within an image. Traditional methods may consider the entire image or multiple regions, which can lead to suboptimal exposure settings when only a specific area is critical. The invention improves upon this by ensuring that exposure values are determined exclusively based on the ROI, ignoring other areas of the image. This approach enhances precision in exposure control, particularly in applications where only a specific portion of the image requires accurate exposure, such as medical imaging, surveillance, or scientific photography. The method involves analyzing the ROI to derive exposure parameters like brightness, contrast, or dynamic range, without influence from surrounding regions. This ensures that the exposure settings are optimized for the critical area, improving image quality and reducing the need for post-processing adjustments. The invention is particularly useful in scenarios where the ROI is small or has distinct lighting conditions compared to the rest of the image. By focusing solely on the ROI, the method avoids distortions or inaccuracies that may arise from averaging or integrating data from irrelevant regions. This targeted approach enhances the reliability and consistency of exposure settings in digital imaging systems.
3. The method of claim 2 , further comprising: determining a separate exposure value from an entirety of the preview image, wherein the exposure value is to be used for capturing an image other than the number of images; receiving the number of images and the image other than the number of images captured using corresponding exposure values; and generating the HDR image based on the number of images and the image other than the number of images.
This invention relates to high dynamic range (HDR) imaging, specifically improving the capture and processing of HDR images by incorporating an additional image with a distinct exposure value. The problem addressed is the limited dynamic range of standard imaging sensors, which often results in either overexposed highlights or underexposed shadows in a single exposure. Traditional HDR techniques combine multiple images with different exposure values to create a final HDR image, but these methods may not fully optimize the dynamic range or may introduce artifacts. The invention enhances HDR imaging by capturing a set of images with varying exposure values and an additional image with a separate exposure value derived from the entire preview image. This additional image is captured independently of the initial set and is used alongside the other images to generate the final HDR image. The exposure value for this additional image is determined based on the entire preview image, ensuring optimal exposure for regions that may not be adequately captured by the initial set of images. The final HDR image is then generated by combining the multiple images, including the additional image, to produce a high-quality result with improved dynamic range and reduced artifacts. This approach ensures better exposure balance across the entire scene, particularly in challenging lighting conditions.
4. The method of claim 1 , further comprising: receiving a user input for the preview image; and identifying one or more ROIs of the preview image from the user input.
This invention relates to image processing, specifically techniques for identifying regions of interest (ROIs) in preview images. The problem addressed is the need for efficient and accurate ROI detection in digital images, particularly when user input is involved. The method involves capturing a preview image from a camera or other imaging device. The preview image is then processed to generate a preview image data set, which includes pixel data and metadata such as exposure settings. The method further includes receiving user input that specifies one or more regions of interest within the preview image. Based on this input, the system identifies and isolates the ROIs for further processing, such as enhanced focus, exposure adjustment, or object tracking. The user input may be provided through touch gestures, voice commands, or other interaction methods, allowing dynamic selection of areas to prioritize in subsequent image capture or analysis. This approach improves image quality by ensuring critical regions are optimized while reducing computational overhead by focusing processing on relevant areas. The method may also be applied in real-time applications like surveillance, medical imaging, or augmented reality, where rapid and precise ROI detection is essential.
5. The method of claim 1 , further comprising for each ROI corresponding to an identified face of the one or more faces: comparing a size of the identified face to a size of the ROI corresponding to the identified face.
This invention relates to image processing, specifically to analyzing regions of interest (ROIs) in images containing one or more faces. The problem addressed is ensuring accurate alignment and sizing of ROIs with detected faces, which is critical for applications like facial recognition, surveillance, and biometric analysis. The method involves detecting one or more faces in an image and identifying a region of interest (ROI) corresponding to each detected face. For each identified face, the method compares the size of the face to the size of its corresponding ROI. This comparison helps determine whether the ROI is appropriately sized relative to the face, ensuring that the ROI captures the entire face without excessive background or missing facial features. The method may also adjust the ROI based on this comparison to improve accuracy in subsequent processing steps, such as facial recognition or feature extraction. The invention improves upon prior art by dynamically verifying and refining ROI sizing, which enhances the reliability of face-related analyses. This is particularly useful in scenarios where faces may vary in size, orientation, or distance from the camera, ensuring consistent performance across different imaging conditions.
6. A device configured to generate a High-Dynamic Range (HDR) image, comprising: one or more processors implemented in circuitry; and a memory coupled to the one or more processors and including instructions that, when executed by the one or more processors, cause the device to: identify one or more faces in a preview image; identify a region of interest (ROI) for each identified face of the one or more faces in the preview image, wherein a number of identified ROIs is more than one; determine a number of images to be captured for an HDR image to be the number of identified ROIs; and for each ROI corresponding to an identified face of the one or more faces: determine a skin tone of the identified face; determine, based on the determined skin tone and a comparison between the identified face and the ROI corresponding to the identified face, a target brightness of the respective image associated with the ROI; and determine, based on the target brightness, an exposure value of the respective image associated with the ROI.
This invention relates to High-Dynamic Range (HDR) imaging systems designed to optimize facial exposure in captured images. The problem addressed is the difficulty in balancing exposure across multiple faces in a scene, particularly when skin tones vary, leading to overexposed or underexposed facial regions in traditional HDR techniques. The device includes processors and memory with instructions to analyze a preview image for faces and define a region of interest (ROI) for each detected face. The number of images captured for HDR processing is set equal to the number of ROIs. For each face, the system determines skin tone and compares the face to its corresponding ROI to calculate a target brightness. Based on this target brightness, an exposure value is assigned to the image associated with that ROI. This ensures that each face is properly exposed in the final HDR composite, accounting for variations in skin tone and lighting conditions. The approach dynamically adjusts exposure settings per face, improving facial clarity and detail in HDR images.
7. The device of claim 6 , wherein execution of the instructions further causes the device to exclusively use the ROI in determining the exposure value.
A system for image capture processes regions of interest (ROIs) within a scene to determine optimal exposure settings. The system identifies one or more ROIs in an image frame, where each ROI represents a specific area of interest for exposure calculation. The system then analyzes the ROIs to determine exposure values, such as shutter speed, aperture, or ISO, based on the characteristics of these regions. The system can prioritize certain ROIs over others, allowing for selective exposure adjustments. In some implementations, the system exclusively uses the ROI data to calculate the exposure value, ignoring other areas of the image. This ensures that the exposure is optimized for the most critical parts of the scene, improving image quality in challenging lighting conditions. The system may also adjust exposure settings in real-time during continuous image capture, such as video recording, to maintain consistent exposure across frames. The ROI-based exposure control can be applied in various imaging devices, including cameras, smartphones, and surveillance systems, to enhance dynamic range and detail preservation.
8. The device of claim 7 , wherein execution of the instructions further causes the device to: determine a separate exposure value from an entirety of the preview image, wherein the exposure value is to be used for capturing an image other than the number of images; receive the number of images and the image other than the number of images captured using corresponding exposure values; and generate the HDR image based on the number of images and the image other than the number of images.
This invention relates to high dynamic range (HDR) imaging systems that capture multiple images with different exposure values to create a final HDR image. The problem addressed is improving HDR image quality by incorporating an additional image captured with a separate exposure value derived from the entire preview image, rather than relying solely on a predefined set of exposures. The system captures a number of images using a set of exposure values and an additional image using a distinct exposure value determined from the preview image. The preview image is analyzed to compute an optimal exposure value for the additional image, which is then captured alongside the other images. The final HDR image is generated by combining all captured images, including the additional one, to enhance dynamic range and detail in both bright and dark regions. This approach ensures better adaptation to varying lighting conditions by dynamically adjusting exposure settings based on real-time preview data, resulting in a more balanced and detailed HDR output. The system is particularly useful in cameras and imaging devices where automatic exposure control is critical for high-quality HDR photography.
9. The device of claim 8 , further comprising one or more cameras, wherein execution of the instructions further causes the device to: capture the preview image using at least one camera of the one or more cameras; and capture the number of images and the image other than the number of images using at least one camera of the one or more cameras.
A device for capturing and processing images includes one or more cameras and a processor executing instructions to perform image capture and analysis. The device captures a preview image using at least one camera and subsequently captures a specified number of additional images and at least one other image using the same or different cameras. The captured images may be used for various applications, such as object recognition, scene analysis, or quality control. The device ensures synchronized or sequential capture of multiple images to enable comparison, enhancement, or further processing. The system may include additional components, such as sensors or memory, to support image storage, transmission, or real-time analysis. The device is designed to improve image acquisition efficiency and accuracy in applications requiring multiple image inputs.
10. The device of claim 6 , further comprising: a display to display the preview image; and a user interface to receive a user input for the displayed preview image; wherein execution of the instructions further causes the device to identify one or more ROIs of the preview image from the user input.
This invention relates to image processing systems that enhance user interaction with preview images, particularly in medical or diagnostic applications. The problem addressed is the need for precise and efficient identification of regions of interest (ROIs) within preview images, which is critical for accurate analysis and diagnosis. The system includes a device with a display to show a preview image and a user interface to receive user input for selecting or marking specific areas within the image. The device processes this input to identify and extract one or more ROIs from the preview image. The user interface allows for interactive adjustments, ensuring that the selected ROIs accurately reflect the areas of interest for further analysis. The system may also include image capture components, such as a camera or scanner, to generate the preview image, and processing logic to analyze the ROIs. The invention improves workflow efficiency by enabling real-time user interaction and precise ROI selection, reducing the need for manual adjustments and enhancing diagnostic accuracy. This is particularly useful in fields like radiology, pathology, or industrial inspection where accurate ROI identification is essential.
11. The device of claim 6 , wherein execution of the instructions, for each ROI corresponding to an identified face of the one or more faces, further causes the device to: compare a size of the identified face to a size of the ROI corresponding to the identified face.
This invention relates to a system for analyzing regions of interest (ROIs) in images, particularly for facial recognition or analysis. The system processes an image to identify one or more faces and then generates ROIs corresponding to each identified face. For each ROI, the system compares the size of the identified face to the size of the ROI. This comparison helps determine whether the ROI accurately captures the face, ensuring proper alignment and scaling for subsequent analysis, such as facial recognition, emotion detection, or identity verification. The system may adjust the ROI based on this comparison to improve accuracy. The invention is part of a broader system that includes image capture, face detection, and ROI generation, ensuring that the analysis is performed on correctly sized and positioned regions. This approach enhances the reliability of facial analysis by minimizing errors caused by misaligned or improperly scaled ROIs. The system may be implemented in software, hardware, or a combination thereof, and can be used in applications such as security systems, biometric authentication, or social media analysis.
12. A non-transitory computer-readable medium storing one or more programs containing instructions that, when executed by one or more processors of a device, cause the device to: identify one or more faces in a preview image; identify a region of interest (ROI) for each identified face of the one or more faces in the preview image, wherein a number of ROIs is more than one; determine a number of images to be captured for an HDR image to be the number of identified ROIs; and for each ROI corresponding to an identified face of the one or more faces: determine a skin tone of the identified face; determine, based on the determined skin tone and a comparison between the identified face and the ROI corresponding to the identified face, a target brightness of the respective image associated with the ROI; and determine, based on the target brightness, an exposure value of the respective image associated with the ROI.
This invention relates to high dynamic range (HDR) imaging systems designed to optimize facial exposure in captured images. The problem addressed is the difficulty in balancing exposure for multiple faces in a scene, particularly when skin tones vary, leading to overexposed or underexposed facial regions in traditional HDR techniques. The system processes a preview image to detect one or more faces and identifies a region of interest (ROI) for each face. The number of ROIs determines how many images will be captured to compose the final HDR image. For each face, the system analyzes the skin tone and compares the face to its corresponding ROI to determine a target brightness. Based on this target brightness, an exposure value is calculated for the image associated with that ROI. This ensures that each face is captured with optimal exposure, accounting for variations in skin tone and lighting conditions. The system dynamically adjusts exposure settings for each ROI, improving facial clarity and detail in the final HDR composite. This approach enhances portrait and group photography by automatically adapting to individual facial lighting needs.
13. The non-transitory computer-readable medium of claim 12 , wherein execution of the instructions further causes the device to exclusively use the ROI of the preview image in determining the exposure value.
The invention relates to digital imaging systems, specifically improving exposure control in camera devices by focusing on a region of interest (ROI) within a preview image. Traditional exposure metering often evaluates the entire image, which can lead to poor exposure in critical areas. This invention addresses the problem by dynamically adjusting exposure settings based solely on the ROI, ensuring optimal lighting conditions for the most important part of the scene. The system captures a preview image, identifies the ROI, and calculates an exposure value using only the pixel data within that region. This approach enhances image quality by prioritizing the subject or focal area, particularly in challenging lighting conditions. The method involves analyzing the ROI's brightness, contrast, or other relevant metrics to determine the appropriate exposure settings, such as aperture, shutter speed, or ISO. By excluding irrelevant areas from the exposure calculation, the system avoids over- or underexposure caused by bright backgrounds or dark foregrounds. The invention is implemented via software instructions stored on a non-transitory computer-readable medium, executed by a camera device to process the preview image and adjust exposure parameters accordingly. This solution is particularly useful in photography and videography, where precise exposure control is critical for capturing high-quality images.
14. The non-transitory computer-readable medium of claim 13 , wherein execution of the instructions further causes the device to: determine a separate exposure value from an entirety of the preview image, wherein the exposure value is to be used for capturing an image other than the number of images; receive the number of images and the image other than the number of images captured using corresponding exposure values; and generate the HDR image based on the number of images and the image other than the number of images.
This invention relates to high dynamic range (HDR) imaging, specifically improving the capture and processing of HDR images by incorporating an additional image with a separately determined exposure value. The problem addressed is the limited dynamic range of standard imaging sensors, which often results in either overexposed highlights or underexposed shadows in a single exposure. Traditional HDR techniques use multiple images with different exposure settings to combine into a single HDR image, but this approach may still miss optimal exposure settings for certain scenes. The invention enhances HDR imaging by capturing a set of images with predefined exposure values and an additional image with a separately determined exposure value. The exposure value for the additional image is calculated from the entire preview image, ensuring it complements the exposures of the other images. The device then receives the multiple images, including the additional image captured with the separately determined exposure, and generates the final HDR image by merging all the images. This method improves dynamic range by ensuring the additional image fills gaps in exposure coverage, resulting in a more balanced HDR output with better highlight and shadow detail. The technique is implemented via executable instructions stored on a non-transitory computer-readable medium, enabling integration into digital cameras or imaging software.
15. The non-transitory computer-readable medium of claim 12 , wherein execution of the instructions further causes the device to: receive a user input for the preview image; and identify one or more ROIs of the preview image from the user input.
This invention relates to image processing systems that allow users to interact with preview images to identify regions of interest (ROIs). The technology addresses the challenge of efficiently and accurately selecting specific areas within an image for further analysis or processing, which is critical in applications such as medical imaging, surveillance, and quality control. The system captures a user input, such as a mouse click, touch, or gesture, and processes this input to determine the boundaries of one or more ROIs within the preview image. The identified ROIs can then be used for tasks like zooming, enhancing, or extracting data from those specific areas. The method ensures that the user's selection is accurately translated into precise coordinates or regions, improving the efficiency of subsequent image analysis. This approach reduces the need for manual adjustments and enhances the accuracy of automated image processing workflows. The system may also support multiple ROIs, allowing users to mark several distinct areas for different processing steps. The invention is particularly useful in scenarios where quick and accurate identification of key image regions is essential, such as real-time monitoring or diagnostic applications.
16. The non-transitory computer-readable medium of claim 12 , wherein execution of the instructions further causes the device to: for each ROI corresponding to an identified face of the one or more faces, compare a size of the identified face to a size of the ROI corresponding to the identified face.
This invention relates to computer vision systems for analyzing facial regions of interest (ROIs) in images or video. The problem addressed is accurately determining the relationship between detected faces and their corresponding ROIs, which is critical for applications like facial recognition, surveillance, and biometric analysis. The system processes an image or video frame to identify one or more faces and their associated ROIs. For each identified face, the system compares the size of the face to the size of its corresponding ROI. This comparison helps assess whether the ROI appropriately captures the face, ensuring accurate subsequent processing such as feature extraction or recognition. The system may adjust the ROI based on this comparison to improve detection accuracy. The invention improves upon prior methods by dynamically validating the spatial relationship between faces and their ROIs, reducing errors in facial analysis tasks. This approach is particularly useful in scenarios where faces may be partially occluded or vary in size due to perspective or distance. The system operates on a computing device executing instructions stored on a non-transitory computer-readable medium, ensuring efficient and reliable performance.
Unknown
April 21, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.